Correlation of the acid-sensitivity of polyethylene glycol daunorubicin conjugates with their in vitro antiproliferative activity

Article abstract of DOI:10.1016/j.bmc.2006.02.007

Polyethylene glycol conjugates with linkers of varying acid-sensitivity were prepared by reacting five maleimide derivatives of daunorubicin containing an amide bond (1) or acid-sensitive carboxylic hydrazone bonds (2-5) with α-methoxy-poly(ethylene glyco

Full text of DOI:10.1016/j.bmc.2006.02.007

Bioorganic & Medicinal Chemistry 14 (2006) 4110–4117
Correlation of the acid-sensitivity of polyethylene
glycol daunorubicin conjugates with their in vitro
antiproliferative activity
Paula C. A. Rodrigues,a Thomas Roth,b Heinz H. Fiebig,b Clemens Unger,a
Rolf Mulhauptc and Felix Kratza,*
¨
aTumor Biology Center, Department of Medical Oncology, Clinical Research, Breisacher Straße 117, D-79106 Freiburg, Germany
bOncotest GmbH, Am Flughafen 12-14, D-79108 Freiburg, Germany
cInstitute of Macromolecular Chemistry, University of Freiburg, Stefan-Meier-Straße 21, D-79104 Freiburg, Germany
Received 7 July 2005; revised 1 February 2006; accepted 3 February 2006
Available online 20 March 2006
Abstract—Polyethylene glycol conjugates with linkers of varying acid-sensitivity were prepared by reacting five maleimide deriva-
tives of daunorubicin containing an amide bond (1) or acid-sensitive carboxylic hydrazone bonds (2–5) with a-methoxy-poly(eth-
ylene glycol)-thiopropionic acid amide (MW 20000) or a,x-bis-thiopropionic acid amide poly(ethylene glycol) (MW 20000). The
polymer drug derivatives were designed to release daunorubicin inside the tumor cell by acid-cleavage of the hydrazone bond after
uptake of the conjugate by endocytosis. In subsequent cell culture experiments, the order of antitumor activity of the PEG dauno-
rubicin conjugates correlated with their acid-sensitivity as determined by HPLC (cell lines: BXF T24 bladder carcinoma and LXFL
529L lung cancer cell line; assay: propidium iodide fluorescence assay). The acid-sensitivity of the link between PEG and daunoru-
bicin is therefore an important parameter for in vitro efficacy.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
for macromolecules, that is, endocytosis, which allows
the macromolecular bound drug to be released in intra-
cellular compartments.
In the past 20 years, numerous drug polymer conjugates
with anticancer agents have been developed with the aim
of improving cancer chemotherapy. Passive targeting of
macromolecules to solid tumors is mediated by the path-
ophysiology of tumor tissue, characterized by a high
metabolic turnover, angiogenesis, a defective vascular
architecture, and an impaired lymphatic drainage.1
In our initial work on acid-sensitive drug conjugates,
we developed transferrin and albumin conjugates with
the anticancer drugs doxorubicin, daunorubicin, and
chlorambucil which have shown promising in vitro
and in vivo antitumor activity.4–10 In these conjugates,
the drug is linked to the protein through a maleimide
spacer molecule, which incorporates a carboxylic hydra-
zone bond as a predetermined breaking point allowing
the bound drug to be released in the acidic environment
of endosomes and/or lysosomes after cellular uptake of
the conjugate.
One of the key issues which has been addressed for estab-
lishing structure–activity relationships of drug polymer
conjugates is the significance of the chemical bond be-
tween the drug and the polymer. Primarily, two types
of bonds have been investigated in some detail: (a)
acid-cleavable bonds2 and (b) peptide bonds which can
be cleaved by lysosomal enzymes.3 Essentially, both
types of bonds exploit the cellular uptake mechanism
Recently, we extended our therapeutic approach to the
synthetic polymer polyethylene glycol (PEG) and devel-
oped PEG conjugates with doxorubicin, paclitaxel, and
methotrexate.11–13 High-molecular weight PEGs are
non-ionic, water-soluble synthetic polymers which are
potential drug carriers due to their synthetic diversity,
their recognized biocompatability, and tumor
accumulation.14
Keywords: Daunorubicin; Polyethylene glycol; Drug polymer conju-
gates; Acid-sensitivity; In vitro activity.
*
Corresponding author. Tel.: +49 761 2062930; fax: +49 761
2062905; e-mail: felix@tumorbio.uni-freiburg.de
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.02.007

P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
4111
In this paper, we report on the synthesis and antiprolif-
erative activity of daunorubicin conjugates with PEGs
of MW 20 kDa which contain an amide bond or hydra-
zone bonds of varying acid-sensitivity. The main objec-
tive of the present work was to investigate whether the
acid-lability of these conjugates correlated with their
inhibitory effects in tumor cell lines, an issue which has
not been addressed to date.
2-amino-4,5-difluorobenzoic acid was reacted with tert-
butyl carbazate in the presence of the coupling reagent
N,N0-dicyclohexylcarbodiimide (DCC) to yield a1 and
a2, respectively. The maleimide derivatives b1 and b2
were then obtained by reaction of 3-maleimidobenzoic
acid chloride with a1 or a2 with 1 equiv of triethylamine
and isolating the compounds through chromatography
on a silica gel column. In a next step, the tert-butoxycar-
bonyl group was cleaved with CF3COOH to yield the
hydrazides c1 and c2, respectively. Finally, the carboxylic
hydrazone derivatives 4 and 5 were obtained by reacting
daunorubicin HCl with an excess of c1 or c2 in anhy-
drous methanol and isolating the products through crys-
tallization. All newly synthesized compounds were
characterized by NMR spectroscopy and mass spec-
trometry (see Section 3). NMR spectra in d6-DMSO
indicated the presence of E(Z)-isomers in this solvent.
2. Results and discussion
2.1. Synthesis of maleimide derivatives of daunorubicin
The five maleimide derivatives of daunorubicin (1–5)
which were coupled to sulfhydryl containing PEGs are
shown in Figure 1. Compounds 1–5 differ in the site
(30-amino or 13-keto position) and stability (benzamide
or substituted as well as non-substituted benzoyl and
phenylacetyl hydrazone bonds) of the chemical link be-
tween daunorubicin and the spacer molecule. Com-
pounds 1, 2, and 3 were synthesized as described
previously.10 The route of preparing 4 and 5 is depicted
in Scheme 1. In a first step, 2-aminobenzoic acid or
2.2. Preparation of PEG daunorubicin conjugates
PEG20000-1 to PEG20000-5 and PEG20000-(1)2 to
PEG20000-(5)2
Polyethylene glycol conjugates of daunorubicin were
prepared by reacting a twofold excess of 1–5 with
O
O
OH
OH
O
CH3
N
O
OH
O
C
OCH3
O
NH
O
OH
OH
N
O
O
CH3
CH3
OH
HO
NH
O
N
O
O
OCH3 O
O
O
2
CH3
HO
1
NH2 HCl
O
O
N
O
C
O
HN
R
R
N
O
O
C
C
O
NH
NH
O
OH
N
O
OH
N
CH3
CH3
OH
OH
4 (R = H)
5 (R = F)
OCH3 O
OH
OCH3 O
OH
O
O
O
O
CH3
CH3
HO
HO
NH2 HCl
NH2 HCl
3
Figure 1. Structures of the maleimide derivatives of daunorubicin 1–5.

4112
P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
O
O
H
N
H
N
6
3
6
19
1
R
O
5
R
R
COOH
NH2
R
R
5
4
1
2
O
8
7
a
7
N
H
b
N
H
20
9
O
O
4
2
R
NH
8
NH2
3
14
9
R = H (b1), R = F (b2)
R = H (a1), R = F (a2)
13
12
O
10
11
R
c
N
15
18
R
O
O
21
H
N
27
24
15
18
17 16
O
OH
O
26
N
O
17
6
13
29
O
HN 22
R
R
NH2•CF3COOH
1
5
4
12
5
7
1
4
11
6
10
7
CH3
N
N
H
d
14
30
32
OH
O
2
NH
8
31
R = H (c1), R = F (c3)
O
3
14
OCH3 O
OH
9
13
12
O
R = H (4), R = F (5)
5'
O
O
10
CH3
4'
1'
11
HO
2'
N
18
15
O
O
NH2 HCl
16
17
Scheme 1. Reagents: (a) H2NHN-COO-C(CH3)3, DMAP, DCC, THF; (b) 3-maleimidobenzoic acid chloride, Et3N, THF; (c) 1—CF3COOH,
2—Et2O (d) daunorubicin HCl, MeOH.
a-methoxy-poly(ethylene glycol)-thiopropionic acid
amide (MW 20,000) or a,x-bis-thiopropionic acid
amide poly(ethylene glycol) (MW 20,000) in aqueous
media. The HS-group in the polymer adds to the
double bond of the maleimide group in a fast and
selective reaction forming a stable thioether bond.
Subsequently, the resulting PEG daunorubicin conju-
gates were isolated through size-exclusion chromato-
graphy over Sephadexꢁ G-25 in phosphate buffer or
over LH20ꢁ in methanol for NMR studies (struc-
tures of the synthesized conjugates are shown in
Fig. 2).
present in the spectra, indicating that the HS-group had
reacted with the maleimide group. Assignment of the
proton signals of the sugar ring of daunorubicin was
not possible, however, due to a very broad signal of
–CH2-groups of the polymer despite our decoupling
attempts.
2.3. pH-dependent stability studies
Previous stability studies have demonstrated an acid-sen-
sitive character of the carboxylic hydrazone bond.4–10
In order to assess distinct differences between the stability
of individual PEG daunorubicin conjugates, we per-
formed pH-dependent stability studies with the conju-
gates at pH 5.0 and 7.4 on our Nucleogelꢁ column with
the aid of HPLC.The decrease in the peak area of the
conjugate recorded at k = 495 nm was used as a measure
of daunorubicin release.
The purity of the samples after gel filtration was deter-
mined with an analytical HPLC-size-exclusion column
(Nucleogelꢁ aqua-OH 40-8) and showed no free dauno-
rubicin or unreacted maleimide daunorubicin deriva-
tives. Retention times of the daunorubicin PEG
conjugates are between 7 and 10 min on this column,
free daunorubicin elutes as a peak at ꢀ70 min (data
not shown). UV/vis-spectra of the daunorubicin PEG
conjugates in phosphate buffer showed the typical
absorption maxima at k = 496, 475, 290, and 234 nm
(data not shown). Selected vacuum-dried samples
[PEG20000-5 and PEG20000-(1)2] which were obtained
over Sephadexꢁ LH20 in methanol were investigated
Whereas the amide derivatives PEG20000-1 did not
release daunorubicin at pH 5.0 or 7.4 after 48 h,
the conjugates containing a carboxylic hydrazone
bond showed good stability at pH 7.4 (less than
10% release of daunorubicin after 48 h) but released
daunorubicin at pH 5.0 in the order PEG-2 > PEG-
4 > PEG-3 ꢁ PEG-5 with half-lives ranging from 7
to >72 h (see Table 1). No noteworthy difference in
acid-sensitivity was observed between the PEG dau-
norubicin conjugates which had one daunorubicin
molecule or two daunorubicin molecules bound to
the polymer. Introduction of the fluor substituents
in the benzoyl hydrazone ring had a pronounced
effect on the acid-sensitivity of the PEG conjugates
(see Table 1). PEG20000-5 and PEG20000-(5)2
showed only marginal acid-sensitivity at pH 5.0.
When the pH was lowered to pH 3.5, half-lives were
of the order of ꢀ20 h.
1
with H NMR spectroscopy in CDCl3 (400 MHz). Eth-
ylene signals of the polyethylene glycol backbone were
decoupled at 3.5 ppm. Analysis of the spectra revealed
that distinct signals of the anthraquinone ring could be
assigned, that is, the HO-6 (ꢀ14 ppm, s) and HO-14
(ꢀ13.3 ppm, s) protons as well as the aromatic protons
of ring A together with two proton signals of the spacer
(7.5–8.0 ppm). In addition, the NH-proton of the car-
boxylic hydrazone bond showed a characteristic peak
at ꢀ10.5 ppm. The characteristic strong singlet signal
of the maleimide double bond (ꢀ7.2 ppm) was no longer

P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
4113
Figure 2. Structures of PEG daunorubicin conjugates PEG20000-1 to PEG20000-5 and PEG20000-(1)2 to PEG20000-(5)2.
2.4. Cell culture experiments
(BXF T24 bladder carcinoma and LXFL 529 lung can-
cer cells) using the propidium iodide fluorescence assay.
Respective IC50 values are summarized in Table 1. Un-
bound polyethylene glycols had only marginal influence
on cell growth in both cell lines (data not shown). The
The newly synthesized daunorubicin PEG conjugates
and unbound daunorubicin were subsequently evaluated
for inhibitory effects in two human tumor cell lines

4114
P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
Table 1. In vitro antitumor activity and half-lives at pH 5.0 of
daunorubicin PEG conjugates in two human tumor cell lines (bladder
carcinoma T24, lung carcinoma LXFL 529)a
acid-cleavage of a predetermined breaking point allow-
ing the drug to be released inside the tumor cell. Our
fluorescence microscopy studies with acid-sensitive
doxorubicin PEG conjugates have shown that doxorubi-
cin is primarily accumulated in the cytoplasm of tumor
cells in contrast to the nucleus which is the predominant
organelle that shows fluorescence after cell exposure to
free doxorubicin.11
Compound
IC50
(LXFL 529L)
(lM)
IC50
(BXF T24)
(lM)
t1/2 (pH 5.0)
(h)
Daunorubicin
PEG20000-1
PEG20000-2
PEG20000-3
PEG20000-4
PEG20000-5
0.008
>10
0.2
0.08
>10
0.04
0.5
—
>72
ꢀ7
0.3
0.5
ꢀ10
ꢀ22
>72b
The potential clinical significance of high-molecular
weight acid-sensitive anthracycline prodrugs is currently
being addressed: the (6-maleimidocaproyl)hydrazone
derivatives of doxorubicin, used as an albumin-binding
prodrug or coupled to the BR96-antibody, are acid-sen-
sitive prodrugs of doxorubicin that are undergoing
phase I/II studies.15–18
1.2
8.8
4.0
PEG20000-(1)2
PEG20000-(2)2
PEG20000-(3)2
PEG20000-(4)2
PEG20000-(5)2
>10
0.06
0.08
0.4
>10
0.2
0.2
0.6
4.9
>72
ꢀ7
ꢀ12
ꢀ24
>72b
2.8
a Similar results were obtained in a second experiment.
b Half-lives at pH 3.5 were ꢀ20 h.
3. Experimental
conjugates PEG-2 and PEG-3 which contain a benzoyl
hydrazone bond in meta-position to the maleimide
group and a phenylacetyl hydrazone bond in para-posi-
tion to the maleimide group are the most active conju-
gates followed by PEG-4 which contains a benzoyl
hydrazone bond in ortho-position to the maleimide
spacer. The difference in the IC50 values between PEG-
2 and PEG-4 which both contain a benzoyl hydrazone
bond can be best explained by the different substitution
position of the maleimide moiety in the aromatic ring
and/or the different nature of the attached chemical
group (maleimide group versus maleimide spacer bound
through an amide bond to the phenyl ring). The de-
crease in the IC50 values between PEG-2, PEG-3, and
PEG-4 correlates with the increase in the half-lives of
the conjugates at pH 5.0. PEG-5, which is analogous
to PEG-4 but contains 2 fluorine atoms in the benzoyl
hydrazone ring, is significantly less active than the other
acid-sensitive conjugates; this observation is in line with
a considerable decrease in the acid-sensitivity of
PEG20000-5 and PEG20000-(5)2 which is attributed to
the electron-withdrawing effect on the benzoyl hydra-
zone bond (see Table 1). The amide conjugate PEG-1
showed no activity at the concentrations tested (0.001–
10 lM). The PEG daunorubicin conjugates which had
two daunorubicin molecules bound to the polymer were
generally more active than the congeners that contain
one daunorubicin molecule. Free daunorubicin is the
most active compound in both cell lines.
3.1. Chemicals, materials, and spectroscopy
1H NMR and 13C NMR: Bruker 400 MHz AM 400 or
Varian Unity 300 (internal standard: TMS); 13C NMR
spectra were obtained with broad-band decoupling;
mass spectra were obtained on a Finnigan MAT 312
with associated MAT SS 200 data system using electron
impact or electro spray ionization; analytical HPLC:
HPLC studies were performed on an analytical HPLC
column (Nucleogelꢁ aqua-OH 40-8, 300 mm · 7.7 mm,
from Macherey & Nagel, FRG); mobile phase: 0.15 M
NaCl, 0.01 M sodium phosphate, 10% v/v CH3CN,
and 30% v/v MeOH, pH 7.0. A Lambda 1000 UV/vis
monitor from Bischoff (at k = 495 or 280 nm), an auto-
sampler Merck Hitachi AS400, and an Integrator Merck
Hitachi D2500 were used, silica gel chromatography on
silica gel 60 (0.063–0.100 mm) from Merck AG; TLC:
silica-coated plates 60 F254 from Merck AG; daunorubi-
cin HCl was a gift from Aventis, FRG; organic solvents:
HPLC grade (Merck) or analytical grade (gift from
BASF AG)—other organic or inorganic compounds:
Merck AG, FRG. Compounds 1, 2, and 3 were pre-
pared previously.10 PEGs were purchased from Rapp
Polymer, FRG; the buffers used were vacuum-filtered
through a 0.2 lm membrane (Sartorius, FRG). Cell cul-
ture media, supplements (L-glutamine, antibiotics, and
trypsin versene/EDTA), and fetal calf serum (FCS) were
purchased from Bio Whittaker (Serva, Heidelberg,
FRG). Propidium iodide was purchased from Aldrich–
Sigma-Chemie, FRG. All culture flasks were obtained
from Greiner Labortechnik (Frickenhausen, FRG).
In summary, we have addressed the relationship be-
tween the pH-dependent stability of the bond between
daunorubicin and PEG in five PEG daunorubicin conju-
gates and their in vitro antitumor activity against two
tumor cell lines. We observed a consistent correlation
between the acid-sensitivity of the conjugates and their
antiproliferative effects, increasing acid-sensitivity being
paralleled by enhanced cytotoxicity. Macromolecular
prodrugs with PEG, like other drug polymer conjugates,
are taken up by the tumor cell through endocytosis.3,7,11
During internalization of the conjugate, the pH is re-
duced from 7.4 to 5.0 in endosomes and pH 4.0 in lyso-
somes, and this pH change can be exploited through
3.2. Methods for the preparation of conjugates
FPLC for preparation of conjugates: P-500 pump, LCC
501 Controller (Pharmacia), and LKB 2151 UV-moni-
tor (at k = 280 nm); buffer: 0.004 M sodium phosphate,
0.15 M NaCl, pH 7.4.
3.3. Synthesis of 4 and 5
Compounds a1 and a2: 17.3 mmol 2-amino-4,5-difluoro-
benzoic acid or 2-aminobenzoic acid, 9.15 g (69.2 mmol,

P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
4115
4 equiv) tert-butylcarbazate and 21.2 mg (0.17 mmol)
DMAP were dissolved in 50 ml THF to which a solution
of 3.93 g (19.0 mmol, 1.1 equiv) N,N0-dicylohexylcarbo-
diimide in 30 ml THF was added during a period of
90 min at 4 ꢂC. The mixture was stirred for 12 h at
4 ꢂC, filtered, and the solvent evaporated in vacuo.
The residue was dissolved in 100 ml ethyl acetate and
extracted six times with 50 ml H2O. After drying over
MgSO4, the organic layer was concentrated to a volume
of 50 ml and n-hexane added until a slight turbidity ap-
peared. The resulting suspension was allowed to stand at
ꢂ20 ꢂC for 12 h to afford a1 and a2 as white crystals
which were washed with diethyl ether and dried under
high vacuum.
CH3)—MS (ESI: Spray 3.8 kV, 200 ꢂC, Sheatgas
30 psi): m/z = 473 (75) [M+• + Na+], 451 (10)
[M+ + 1]ꢂ(C23H22N4O6) M = 450.45 g/mol.
Compound b2: yield 3.5 g (7.2 mmol, 51.4%)—1H NMR
(300 MHz, DMSO-d6/TMS): d = 10.9 (s, 1H, C7-NH),
7.6 (dd, 1H, 6-H, 3J6-H,5-F = 11.9 Hz, 4J6-H,4-F = 9.2 Hz),
7.64- 7.55 (m, 4H, arom. H), 7.2 (s, 2H, 16-H/17-H), 6.8
(dd, 1H, 3-H, 3J3-H,4-F = 13.7 Hz, 4J3-H,5-F = 7.8 Hz), 6.7
(br s, 2H, C2-NH, C19-NH), 1.2 (s, 9H, 3 CH3)—13C
NMR (75.4 MHz, DMSO-d6/TMS): d = 169.8 (C-15/C-
1
18), 166.3 (C-8), 151.1 (C-4, J(4-C/4-F) = 249.5 Hz),
1
149.2 (C-7), 148.9 (C-19), 140.1 (C-5, J(5-C/5-F)
=
233.4 Hz), 136.8 (C-2), 136.7/131.9/130.0/129.2/126.8/
125.8 (arom. C), 134.9 (C-16/C-17), 116.5 (C-1), 105.9
(C-6, J(6-C/5-F) = 14.1 Hz), 104.0 (C-3, J(3-C/4-F) =
2
2
Analytical data. Compound a1: yield 3.73 g (20.2 mmol,
40%)—1H NMR (300 MHz, MSO-d6/TMS): d = 9.8/8.8
(2s, 2H, NH–NH), 7.6 (d, 1H, 6-H), 7.2 (t, 1H, 4-H),
6.8 (d, 1H, 5-H), 6.6 (t, 1H, 3-H), 6.4 (br s, 2H, NH2),
1.4 (s, 9H, 3 CH3)—13C NMR (75.4 MHz, DMSO-
d6/TMS): d = 168.4 (C-7), 156.4 (C-8), 149.8 (C-6),
132.1 (C-1), 127.9 (C-4), 116.3 (C-2), 114.4 (C-5),
112.2 (C-3), 79.9 (C-9), 28.0 (3 CH3) MS (EI, 220 ꢂC,
70 eV, 1 mA, 2 kV): m/z = 251(18) [M+•], 196
(33) [(M+1)+•ꢂC4H8+], 147 (32) [(M+1)+•ꢂ(CO2,
–C4H8+)]—(C12H17N3O3) M = 251.28 g/mol.
14.1 Hz), 84.1 (C-9), 27.2 (3 CH3)—MS (ESI: Spray
4-5 kV, 200–250 ꢂC, Sheatgas 20 psi): m/z = 471 (100)
[M+•ꢂCH3+], 439 (56) [M+•ꢂ(CO, F+)], 423 (66)
[M+ꢂ(CO2, F+)], 389 (24) [M+•ꢂC4H2NO2+]—
(C23H20F2N4O6) (486.4 g/mol).
Compounds c1 and c2: 2.7 mmol b1 or b2 was suspended
in 3 ml trifluoroacetic acid and stirred at room tempera-
ture until all of the compound was dissolved. Excess of
trifluoroacetic acid was evaporated in vacuo and the oily
residue triturated with diethyl ether to afford c1 and c2 as
a yellow powder.
Compound a2: yield 3.85 g (13.4 mmol, 78%)—1H
NMR (300 MHz, DMSO-d6/TMS): d = 9.9/8.8 (2s, 2H,
NH-NH), 7.6 (dd, 1H,. 6-H, J6-H,5-F = 11.9 Hz, J6-
3
4
Analytical data. Compound c1: yield 1.17 g (2.5 mmol,
96%)—1H NMR (300 MHz, DMSO-d6/TMS): d = 11.6
(br s, 1H, NH2ÆCF3COOH), 8.4 (s, 1H, C7-NH), 8.0–
7.3 (m, 5H, 4 arom. H, C8-NH), 7.21 (s, 2H, 16-H/17-
H)—13C NMR (75.4 MHz, DMSO-d6/TMS): d = 169.8
(C-15/C-18), 167.5 (C-7), 164.0 (C-8), 138.2/135.5/
133.0/132.3/130.3/129.4/128.5/126.0/125.3/123.9/122.2/
120.4 (arom. C), 135.2 (C-16/C-17), 123.8/118.0/ 114.1/
111.5 (CF3)—MS (ESI: Spray 3.8 kV, 200 ꢂC, Sheatgas
30 psi): m/z = 351 (100) [M+•ꢂCF3COOꢂ], 319 (58)
3
H,4-F = 9.2 Hz), 6.7 (dd, 1H, 3-H, J3-H,4-F = 13.2 Hz,
4J3-H,5-F = 7.2 Hz), 6.6 (br s, 2H, NH2), 1.4 (s, 9H, 3
CH3)—13C NMR (DMSO-d6/TMS): d = 166.7 (C-7),
155.5 (C-8), 152.1 (C-4, J(4-C/4-F) = 249.5 Hz), 148.3
1
1
(C-5, J(5-C/5-F) = 233.4 Hz), 139.7 (C-2), 116.1 (C-1),
2
2
107.1 (C-6, J(6-C/5-F) = 14.1 Hz), 103.48 (C-3, J(3-C/4-
F) = 14.1 Hz), 79.2 (C-9), 28.0 (3 CH3)—MS (EI,
220 ꢂC, 70 eV, 1 mA, 2 kV): m/z = 287 (18) [M+•], 231
(67) [(M + 1)+•ꢂC4H8+], 187 (32) [(M + 1)+•ꢂ(CO2,
–C4H8+)],
(C12H15F2N3O3) M = 287.27 g/mol.
156
(100)
[M+•ꢂ(NH–NH–C4H9)+],
[M+ꢂNH-NH3CF3COOꢂ]—(C20H13F5N4O6)
M =
464.35 g/mol.
Compounds b1 and b2: 3.63 g (15.4 mol, 1.1 equiv) 3-
maleimidobenzoic acid chloride dissolved in 90 ml
THF was added during a period of 90 min at room tem-
perature to a stirring solution of 14 mmol a1 or a2 and
2.15 ml (15.4 mmol, 1.1 equiv) triethylamine in 140 ml
THF. The reaction mixture was stirred for 72 h, the trie-
thylammonium chloride salt removed by filtration, and
the mixture concentrated to an oily residue. Purification
was performed by chromatography (silica gel in ethyl
acetate/hexane, 1.5:1) to yield b1 and b2 as yellow
crystals.
Compound c2: yield 0.98 g (2 mmol, 75%)—mp
192 ꢂC; 1H NMR (300 MHz, DMSO-d6/TMS):
d = 10.47 (s, 1H, C7-NH), 10.2 (br s, 1H, NH2ÆCF3-
COOH), 8.05–7.56 (m, 5H, arom. H, C8-NH), 7.58
3
4
(dd, 1H, 6-H, J6-H,5-F = 11.9 Hz, J6-H,4-F = 9.2 Hz),
=
3
7.21 (s, 2H, 16-H/17-H), 6.7 (dd, 1H, 3-H, J3-H,4-F
4
13.7 Hz, J3-H,5-F = 7.8 Hz)—13C NMR (75.4 MHz,
DMSO-d6/TMS): d = 169.6 (C-15/C-18), 166.7 (C-8),
165.1 (C-7), 152.2 (C-4, J(4-C/4-F) = 249.5 Hz), 148.3
1
1
(C-2), 139.8 (C-5, J(5-C/5-F) = 233.4 Hz), 136.7/131.9/
130.0/129.2/126.8/125.8 (arom. C), 134.9 (C-16/C-17),
116.34 (CF3), 116.24 (C-6, 2J(6-C/5-F) = 14.1 Hz),
105.2 (C-1), 104.5 (C-3, 2J(3-C/4-F) = 14.1 Hz)—
(C20H13F5N4O6) M = 500.33 g/mol.
Analytical data. Compound b1: yield 4.41 g (9.2 mmol,
NMR
70%)—1H
(300 MHz,
DMSO-d6/TMS):
d = 11.9/10.6 (s, 1H, C7-NH), 9.1/8.7 (s,1H, C19-NH),
8.6 (s, 1H, C2-NH), 7.95–7.6 (m, 7H, arom. H), 7.2 (s,
2H, 16-H/17-H), 1.4 (s, 9H, 3 CH3)—13C NMR
(75.4 MHz, DMSO-d6/TMS): d = 169.7 (C-18/C-21),
166.5 (C-7), 163.5 (C-11), 155.2 (C-8), 134.7 (C-19/C-
Compounds 4 and 5: 50 mg (0.089 mmol) daunorubicin
hydrochloride and 0.44 mmol (5 equiv) c1 or c2 were
dissolved in 45 ml anhydrous methanol to which 50 ll tri-
fluoroacetic acid was added. The solution was stirred in
the dark for 96 h at room temperature and then concen-
trated to a volume of approximately 15–20 ml in vacuo.
20), 139.0/132.8/132.4/131.8/131.5/130.8/129.3/128.0/
127.3/125.7/123.2/ 120.7 (arom. C), 79.5 (C-9), 27.9 (3

4116
P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
Acetonitrile was added to the red dark solution until a
slight turbidity appeared. The resulting suspension was
allowed to stand at ꢂ20 ꢂC for crystallization of the prod-
uct. The red hydrazone was collected by centrifugation.
The supernatant was evaporated to a small volume and
treated with acetonitrile as above. The hydrazone frac-
tions were combined and re-crystallized from methanol/
acetonitrile to yield a red microcrystalline powder.
101.7 (C-18), 81.1 (C-9), 80.7 (C*-9), 75.8 (C-7), 68.5
(C-50), 66.2 (C-40), 56.6 (-OCH3), 46.2 (C-30), 38.6 (C-
8), 38.4 (C*-8), 33.9 (C-14), 20.5 (C*-14), 29.4 (C-10),
27.8 (C-20), 16.6 (C-5-CH3) (* = splitting of the proton
or carbon signals that are indicative of the presence of
E(Z)-isomers in DMSO)—MS (ESI: Spray 4–5 kV,
200–250 ꢂC, Sheatgas 20 psi, CH3OH): m/z = 928 (100)
[M + 1 (CH3OH)]+ (methanol adduct), 896 (2)
[M + 1]+, 799 (27) [MꢂC4H2NO2]+— (C45H40F2N5O13)
M = 895.8 g/mol.
Analytical data. Compound 4: yield 60 mg (0.065 mmol,
72%)—1H NMR (400 MHz, DMSO-d6/TMS): d = 14.08
(s, 1H, C-6-OH), 13.30 (s, 1H, C-11-OH), 11.74 (s, 1H,
N-NH), 10.75 (s, 1H, N-NH), 10.21 (s, 1H, C17-NH),
8.21-7.45 (m, 10H, arom. H: 1-H/3-H/2-H/H-18/H-19/
3.4. Synthesis of PEG daunorubicin conjugates
All reactions were performed at room temperature un-
less otherwise stated. Data for one representative exper-
iment are given.
H20/21-H/24-H/26-H/27-H/28-H),
7.86
(m,
1H,
NH2ÆHCl), 7.21 (s, 2H, 30-H/31-H), 7.2 (s, 2H, 30-H/
31-H), 5.49 (m, 1H, 10-H), 5.28 (br s, 1H, C-9-OH),
4.50 (br s, 1H, 7-H), 4.32 (m, 1H, C40-OH), 4.27 (d,
1H, 50-H), 3.96 (s, 3H, –OCH3), 3.61 (m, 1H, 40-H),
2.87 (m, 2H, 10-H), 2.28 (m, 1H, 30-H), 2.17 (m, 2H,
8-H), 2.07 (m, 2H, 20-H), 1.86 (s, 3H, 14-H), 1.7 (s,
3H, 14-H), 1.1 (m, 3H, 50-CH3); 13C NMR (Varian
300, 75.4 MHz, DMSO-d6/TMS): d = 186.3 (C-5),
186.1 (C-12), 169.7 (C-29/C-32), 167.5 (C*-29/C*-32),
163.9 (C-4), 160.6 (C-15), 156.7 (C-11), 156.3 (C-6),
138.3 (C-13), 138.1 (C*-13), 135.1 (C-2), 135.0 (C-6a),
134.7 (C-12a), 134.6 (C-30/C-31), 133.0 (C-10a), 132.2
(C-17), 135.8/131.1/130.2/129.3/128.7/127.4/126.1/125.9/
125.7/123.8/120.3/118.4 (arom. C), 122.0 (C-1), 119.8
(C-4a), 118.9 (C-3), 110.3 (C-5a), 110.0 (C-11a), 109.2
(C-10), 79.5 (C-9), 78.2 (C-9), 76.8 (C-7), 66.2 (C-50),
65.9 (C-40), 56.5 (-OCH3), 46.5 (C-30), 40.2 (C-8), 39.9
(C-8), 38.8 (C-14), 23.4 (C*-14), 28.4 (C-10), 27.2 (C-
20), 16.7 (C-5-CH3) (* = splitting of the proton or carbon
signals that are indicative of the presence of E(Z)-
isomers in DMSO). MS (ESI: Spray 3.8 kV, 200 ꢂC,
Sheatgas 30 psi, CH3OH): m/z = 892 (100) [M + 1
(CH3OH)]+ (methanol adduct), 860 (2) [M+1]+—
(C45H42N5O13) M = 859.8 g/mol.
3.4.1. Preparation of PEG 20000-(2)2. Eight milligrams
(0.01 mmol) of 2 was dissolved in 250 ll dimethylform-
amide and added to 50 mg (0.0025 mmol) PEG-
20000(SH)2 dissolved in 5 ml buffer (0.004 M sodium
phosphate, 0.15 mol NaCl, pH 6.8). The mixture was
homogenized and kept at room temperature for
30 min. After centrifuging the slightly turbid mixture
for 5 min with a Sigma 112 centrifuge, the supernatant
was loaded on
a
Sephadexꢁ
G
25 column
(100 mm · 20 mm, loop size: 5 ml). The conjugate eluted
with a retention time of 5–10 min (flow: 1.0 ml/min,
buffer: 0.004 M sodium phosphate, 0.15 M NaCl, pH
7.4). Concentration of the conjugate to a volume of
approximately 2 ml was carried out with CENTRI-
PREPꢁ-10-concentrators from Amicon, FRG (60 min
at 4 ꢂC and 4500 rpm). The concentration of daunorubi-
cin in the conjugate was adjusted to c = 300 20 lM
using the e-value for daunorubicin in physiological buff-
er (e495 = 9280 Mꢂ1 cmꢂ1)xy. The conjugate was stored
at ꢂ80 ꢂC. Daunorubicin polyethylene glycol conjugates
for NMR studies were chromatographed over
Sephadexꢁ LH20 Gel (100 mm · 10 mm, loop size:
2 ml, flow: 1.0 ml/min, retention time: 3–6 min, eluent:
100% methanol HPLC grade).
Compound 5: yield 60 mg (0.065 mmol, 72%)—1H
NMR (300 MHz, DMSO-d6/TMS): d = 14.08 (s, 1H,
C-6-OH), 13.30 (s, 1H, C-11-OH), 10.92 (s, 1H, N-
NH), 10.30 (s, 1H, N-NH*), 10.21 (s, 1H, C17-H),
8.21/7.94 (m, 2H, 1-H/3-H), 7.86 (m, 1H, NH2ÆHCl),
7.74 (m, 1H, 2-H), 7.68–7.45 (m, 5H, 21-H/24-H/26-
H./27-H/28-H), 7.21 (s, 2H, 30-H/31-H), 7.2 (s, 0 2H,
30-H*/31-H*), 6.68 (dd, 1H, 18-H), 5.49 (m, 1H, 1 -H),
5.28 (br s, 1H, C-9-OH), 4.88 (br s, 1H, C-9-OH*),
4.80 (br s, 1H, 7-H), 4.72 (m, 1H, C40-OH), 4.47 (d,
1H, 50-H), 3.96 (s, 3H, –OCH3), 3.61 (m, 1H, 40-H),
2.87 (m, 2H, 10-H), 2.48 (m, 1H, 30-H), 2.27 (m, 2H,
8-H), 1.90 (m, 2H, 20-H), 1.76 (s, 3H, 14-H), 1.7 (s,
3H, 14-H*), 1.1 (m, 3H, 50-CH3)—13C NMR (Varian
300, DMSO-d6/TMS): d = 186.5 (C-5), 186.4 (C-12),
169.7 (C-29/C-32), 168.3 (C*-29/C*-32), 166.0 (C-22),
160.4 (C-4), 159.3 (C-15), 156.3 (C-11), 154.9 (C-6),
151.9 (C-19), 144.2 (C-13), 144.1 (C*-13), 140.5 (C-20),
136.1 (C-2), 135.5 (C-6a), 135.4 (C-12a), 134.8 (C-30/
C-31), 134.7 (C-10a), 133.2 (C-17), 132.4/131.9/130.2/
3.4.2. pH-dependent stability studies with the PEG
daunorubicin conjugates at pH 5.0 and 7.4. Fifty microli-
ters of the stock solutions of the conjugates
(c 300 20 lM) in phosphate buffer was added to
450 lL of buffer, pH 5.0 (0.15 M NaCl, 0.004 M sodium
phosphate adjusted to pH 5.0 with hydrochloric acid) or
pH 7.4 (0.15 M NaCl, 0.004 M sodium phosphate). The
solutions were incubated at room temperature and
50 lL samples were analyzed at k = 495 nm every 2–
4 h over a period of 72 h on an analytical HPLC column
(Nucleogelꢁ aqua-OH 40-8, 300 mm · 7.7 mm, from
Macherey & Nagel, FRG); mobile phase: 0.15 M NaCl,
0.01 M sodium phosphate, 10% v/v CH3CN, 30% v/v
MeOH, pH 7.0. A Lambda 1000 UV/vis monitor from
Bischoff (at k = 495 nm), an autosampler Merck Hitachi
AS400, and an Integrator Merck Hitachi D2500 were
used.
129.4/127.4/126.1
120.0 (C-1), 119.7 (C-4a), 118.9 (C-3), 115.8 (C-16),
110.5 (C-5a), 110.4 (C-11a), 110.3 (C-10), 108.8 (C-21),
(C-23/C-24/C-25/C-26/C-27/C-28),
3.4.3. Biology. Human tumor cells were grown at 37 ꢂC
in a humidified atmosphere (95% air, 5% CO2) in mono-
layer RPMI 1640 culture medium with phenol red

P. C. A. Rodrigues et al. / Bioorg. Med. Chem. 14 (2006) 4110–4117
4117
supplemented with 10% heat-inactivated FCS, 300 mg/L
References and notes
glutamine, and 1% antibiotic solution (5.000 lg genta-
mycin/mL). Cells were trypsinized and maintained twice
a week. The concentration of free daunorubicin and
daunorubicin in the stock solution of the conjugates
was 300 lM.
1. Maeda, H.; Wu, J.; Sawa, T.; Matsumura, Y.; Hori, K.
J. Control. Release 2000, 65, 271–284.
2. Kratz, F.; Beyer, U.; Schutte, M. T. Crit. Rev. Ther. Drug
¨
Carrier Syst. 1999, 16, 245–288.
3. Duncan, R. Nat. Rev. 2003, 2, 347–359.
4. Kratz, F.; Beyer, U.; Collery, P.; Lechenault, F.; Cazabat,
A., et al. Biol. Pharm. Bull. 1998, 21, 56–61.
3.4.4. Propidium iodide fluorescence assay. The fluores-
cence assay was performed according to the method
of Dengler et al.19 Briefly, cells were harvested from
exponential phase cultures growing in RPMI culture
medium by trypsinization, counted, and plated in
96-well flat-bottomed microtiter plates (50 lL cell sus-
pension/well, 1.0 · 105 cells/mL). After a 24 h recovery
in order to allow cells to resume exponential growth,
100 lL culture medium (6 control wells per plate) or
culture medium containing drug was added to the
wells. Each drug concentration was plated in tripli-
cate. After 6 days of continuous drug exposure, non-
viable cells were stained by addition of 25 lL of a
propidium iodide solution (50 lg/mL). Fluorescence
(FU1) was measured using a Millipore Cytofluor
2350 microplate reader (excitation 530 nm, emission
620 nm). Microplates were then kept at ꢂ18 ꢂC for
24 h, which resulted in a total cell kill. After thawing
of the plates and a second fluorescence measurement
(FU2), the amount of viable cells was calculated by
FU2 ꢂ FU1. Growth inhibition was expressed as treat-
ed/control · 100 (%T/C).
5. Kratz, F.; Beyer, U.; Roth, T.; Tarasova, N.; Collery, P.,
et al. J. Pharm. Sci. 1998, 87, 338–346.
6. Kratz, F.; Roth, T.; Fichtner, I.; Schumacher, P.; Fiebig,
H. H., et al. J. Drug Targeting 2000, 8, 305–318.
7. Beyer, U.; Rothern-Rutishauser, B.; Unger, C.; Wunderli-
Allenspach, H.; Kratz, F. Pharm. Res. 2001, 18, 29–38.
8. Kratz, F.; Beyer, U.; Roth, T.; Schutte, M. T.; Unold, A.,
¨
et al. Arch. Pharm. Pharm. Med. Chem. 1998, 331, 47–53.
9. Beyer, U.; Roth, T.; Schumacher, P.; Maier, G.; Unold,
A., et al. J. Med. Chem. 1998, 41, 2701–2708.
10. Kratz, F.; Beyer, U.; Schumacher, P.; Krueger, M.; Zahn,
H., et al. Bioorg. Med. Chem. Lett. 1997, 7, 617–622.
11. Rodrigues, P. A. C.; Beyer, U.; Schumacher, P.; Roth, T.;
Fiebig, H. H., et al. Bioorg. Med. Chem. 1999, 7, 2517–2524.
12. Rodrigues, P. C. A.; Scheuermann, K.; Stockmar, C.;
Maier, G.; Fiebig, H. H., et al. Bioorg. Med. Chem. Lett.
2003, 13, 355–360.
13. Riebeseel, K.; Biedermann, E.; Lo¨ser, R.; Breiter, N.;
Hanselmann, R., et al. Bioconjug. Chem. 2002, 13, 773–785.
14. Greenwald, R. B.; Conover, C. D.; Choe, Y. H. Crit. Rev.
Ther. Drug Carrier Syst. 2000, 17, 101–161.
15. Dengler, W. A.; Schulte, J.; Berger, D. P.; Mertelsmann,
R.; Fiebig, H. H. Anticancer Drugs 1995, 6, 522–532.
16. Kratz, F.; Warnecke, A.; Scheuermann, K.; Stockmar, C.;
Schwab, J., et al. J. Med. Chem. 2002, 45, 5523–5533.
17. Kratz, F.; Warnecke, A.; Schmid, B.; Chung, D.; Gitzel, M.
Curr. Med. Chem.-Anti-Cancer Agents 2005, 13, 763–771.
18. Gmehling,D.;Medinger,M.;Mross,K.;Haering,B.;Kratz,
F., et al. Proc. Am. Assoc. Cancer Res. 2005, 46, 3987.
19. clinicaltrials@seagen.com.
Acknowledgment
This work has been supported by a grant from the
Fonds der Chemischen Industrie (BMBF).